Show the code

import pandas as pd
import numpy as np
from lets_plot import *

LetsPlot.setup_html(isolated_frame=True)

df = pd.read_csv("https://github.com/byuidatascience/data4names/raw/master/data-raw/names_year/names_year.csv")

Client Request:

The dataset provided contains yearly counts of baby names in the United States. The client, who has a passion for historical naming patterns, would like an analysis of how name usage has changed over time. They are particularly interested in the names Mary, Martha, Peter, and Paul. In addition, they would like to see how the popularity of a name from a well-known movie has shifted following its release.

Summary of overall project work.

Key deliverables included:

Personalization & historical context: Compared the popularity of my own name (Maia) at my birth year against its historical usage.

Generational inference: Examined names like Brittany, identifying sharp peaks (around 1990) that allow us to infer likely age ranges.

Client’s focus names: Analyzed Mary, Martha, Peter, and Paul from 1920–2000, highlighting cultural and religious trends in naming.

Pop culture impact: Investigated names from movies (Elsa from Frozen and Elliot from E.T.), showing clear surges in popularity following major film releases.

This project demonstrates how demographic data, when paired with visualization, can answer client-driven questions, uncover cultural patterns, and tell engaging stories about history and identity.

How does your name at your birth year compare to its use historically?

After filtering the data, I was able to find that in 1999, the name “Maia” was given to a total of 370 babies in the U.S. It remained an relatively uncommon name, but showed some regional popularity clusters in larger more populated stats. The state with the most names was California with 66 babies that shared my name. This accounts for nearly 18% of the national total. I was born in Arizona, which had 8 babies born in 1999 named Maia, making it more meaninful to me to see my name reflected in the data from my birth year and state. Looking at the chart I created, the data suggests that while Maia wasn’t a top national name, it has been steadily adopted across a variety of coast, particularly in the coasts and in urbanized regions. The black line in the chart represents the underlying trend over time, a steady increase in popularity starting around the late 1990s and continuing past 2010. I expect that it will remain at this steady rate unless an outside source influences our society (Ex: popculture, movies, books, etc.).

Show the code

#Package that turns data from python lists, dictionaries, pandas df to nicely formatted tables. clean, printable string
from tabulate import tabulate

#Naming my variable and assigning my year of birth so that I can look up the data within this year
my_name = "Maia"
birth_year = 1999

#The df is shorthand for DataFrame, with this specific application here, it allows me to create a new variable to store the filtered DataFrame, which only contains rows for the name"Maia"

#The .copy() creates a separate copy of the filtered DataFrame to avoid warnings when modifying it later.
my_name_df = df[df['name'] == my_name].copy()

#The "my_name_df...th_year" creates a Boolean filter: Returns TRUE for rows where the year is 1999, and FALSE otherwise.
birth_year_count = my_name_df[my_name_df['year'] == birth_year]

#Prints the filtered DataFrame to show how many babies named "Maia" were born in 1999, across all U.S. states.
print(birth_year_count)

#Creates a new DF named maia_df that includes only the rows where the baby name is "Maia". The .copy() ensures you're working with a safe, modifiable version.
maia_df = df[df['name'] == 'Maia'].copy()

#Helps weed out the potential outliers so that the graphs I create later don't have skewed data based on outliers
potential_outliers = maia_df.query("Total > 500 | (Total > 300 & year > 2010)").copy()

#This part actually shows how many babies were named Maia ineach U.S. state in 1999
maia_1999 = df[(df['name'] == 'Maia') & (df['year'] == 1999)].copy()

#This cleans the data so that the other unnecessary columns don't appear
maia_1999_clean = maia_1999.drop(columns=['name', 'year'])

#This just moves the data from a wide format to a long format in an attempt to make the data easier to read
maia_long = maia_1999_clean.melt(var_name='State', value_name='Count')

#I didn't like that it showed even the states that had no babies named "Maia" so I filtered it to not do that.
maia_nonzero = maia_long[maia_long['Count'] > 0]

#This sorts the data by most popular to least popular, easier to read.
maia_sorted = maia_nonzero.sort_values(by='Count', ascending=False).reset_index(drop=True)

#Prints a title above the table
print("Top States for the Name 'Maia' in 1999:\n")
print(tabulate(maia_sorted, headers='keys', tablefmt='github', showindex=False))

        name  year   AK   AL   AR   AZ    CA   CO    CT   DC  ...   TN    TX  \
252672  Maia  1999  0.0  0.0  0.0  8.0  66.0  0.0  11.0  0.0  ...  0.0  18.0   

         UT    VA   VT    WA   WI   WV   WY  Total  
252672  7.0  15.0  0.0  14.0  5.0  0.0  0.0  370.0  

[1 rows x 54 columns]
Top States for the Name 'Maia' in 1999:

| State   |   Count |
|---------|---------|
| Total   |     370 |
| CA      |      66 |
| NY      |      38 |
| PA      |      22 |
| OH      |      21 |
| TX      |      18 |
| FL      |      16 |
| MA      |      16 |
| VA      |      15 |
| WA      |      14 |
| IL      |      14 |
| MN      |      13 |
| NJ      |      12 |
| OR      |      12 |
| NC      |      11 |
| CT      |      11 |
| MI      |       9 |
| MD      |       8 |
| AZ      |       8 |
| MO      |       7 |
| UT      |       7 |
| NM      |       6 |
| IN      |       6 |
| RI      |       5 |
| IA      |       5 |
| GA      |       5 |
| WI      |       5 |

Show the code

(
    ggplot(maia_df, aes(x='year', y='Total')) +
    #plots each data point (a year w/ a total count for 'Maia') as a black dot
    geom_point(color='black') +
    #smooths out the year-to-year fluctuations and helps underly the trend over time
    geom_smooth(se=False, method='loess', color='black') + 
    #Highlights specific potential outliers
    geom_point(data=potential_outliers, color='red') +
    geom_label(
        aes(label='year'),
        data=potential_outliers,
        color='red',
        position=position_jitter(),
        fontface='bold',
        size=5,
        hjust='left',
        vjust='bottom',
    ) +

    #This creates a vertical line (hint the V line) to show the year that represents my birth, giving it more readability and allows someone to look for the data around my name.
    geom_vline(xintercept=1999, color='blue', linetype='dashed') +
    labs(
        title="Babies named 'Maia': Outlier Years and Trends",
        subtitle="Outliers highlighted in red with labels",
        x="Year",
        y="Total Babies Named Maia"
    ) +
    #Gives chart a clean, uncluttered look for clarity
    theme_minimal() +
    #Remove legend because everything is explained visually
    theme(legend_position='none')
)

If you talked to someone named Brittany on the phone, what is your guess of his or her age? What ages would you not guess?

The popularity of the name ‘Brittany’ peaked in the year 1990. The graph and trend lines show a rapid increas in popularity up to that year, followed by a rapid decline. Based on this data, I would guess that someone on the phone named Brittany would be in their early 30s because the peak year. I would not expect someone significantly younger such as 20 or significantly older over 40 to have this name, since its popularity was heavily concentrated in that narrow time window.

Show the code

#Similar to what I did earlier, filers the dataset to only the rows where the baby name is "Brittany"
#The .copy() lets me work with a clean, independent copy of the data. That way it doesn't use mine from earlier
brittany_df = df[df['name'] == 'Brittany'].copy()

#Filters data by year and give the total of all babies named Brittany across all states for each year
brittany_df = brittany_df.groupby('year', as_index=False)['Total'].sum()

#Finds peak year (using max) that name had highest total count
#.values[0] extracts the actual year number as an integer otherwise it will return as a pandas series
peak_year = brittany_df[brittany_df['Total'] == brittany_df['Total'].max()]['year'].values[0]

#prints output of results
print(f"Brittany peaked in: {peak_year}")

Brittany peaked in: 1990

Show the code

(
    ggplot(brittany_df, aes(x='year', y='Total')) +

    #draws a purple line showing the popularity trend, typing 'purple' didn't seem to work, had to use the color code for purple. I wanted to try a different color besides black
    geom_line(color='#FF1493', size=1.2) +  

    #highlights using smooth line the general trend over time, easier to see rise and fall of name's popularity
    geom_smooth(se=False, method='loess', color='black') +

    #marks the peak year with red dashed line, making it easier to see
    geom_vline(xintercept=peak_year, color='red', linetype='dashed', size=1) +
    labs(
        title="Popularity of the Name 'Brittany' Over Time",
        subtitle=f"Peak year: {peak_year}",
        x="Year",
        y="Total Babies Named Brittany"
    ) +

    #removes chart clutter and emphasizes the data
    theme_minimal()
)

Mary, Martha, Peter, and Paul are all Christian names. From 1920 - 2000, compare the name usage of each of the four names in a single chart. What trends do you notice?

From 1920 - 2000, all four names show unique trends in popularity. This data shows that Mary reached it’s popularity peak in the early to mid-1900s before its steady deciline afters the 60’s. Martha shows a much more mild curve with a slower decline that begins around the 1950’s. Paul had a mid peak but immediatley declined in popularity after the 1970’s The data conveys that Peter is a consistently popular name that peaked during the mid-century and has gradually declined since. These patterns reflect shifts in cultural religious naming trends over the decades. It would be interesting to compare the name Peter with not just the Christian name, but Spider-man movie releases, which I believe would popularize the name Peter (Spider-man’s name is Peter Parker).

Show the code

#Assigns the Christian names to the name variable
names = ['Mary', 'Martha', 'Peter', 'Paul']

#Filters the dataset to pull the four target names, also keeps the search between 1920 - 2000
df_subset = df[df['name'].isin(names) & df['year'].between(1920, 2000)].copy()

#Groups and sums the total column across all states for each name during the outlines years
grouped = df_subset.groupby(['year', 'name'], as_index=False)['Total'].sum()

( #Plots one line per name, colored distinctly
    ggplot(grouped, aes(x='year', y='Total', color='name')) +
    geom_line(size=1.2) +
    labs(
        title="Name Usage of Mary, Martha, Peter, and Paul (1920–2000)",
        subtitle="Christian names over time",
        x="Year",
        y="Total Babies Named",
    ) +
    #Clean and professional appearance
    theme_minimal()
)

Think of a unique name from a famous movie. Plot the usage of that name and see how changes line up with the movie release. Does it look like the movie had an effect on usage?

One of the most popular films to be released in the year 2013 was Disney’s Frozen. After this movie there was a surge in popularity. The chart shows the release year with a dashed black line. An immediate increase following that year reflects how influential the movie was in shaping modern baby name trends. The name did exist before the movie and appeared to have a gradual increase before but it peaked after the movie release. However, after 2015, it began to decline again.

Show the code

# Filter and group data for the name 'Elsa'. The df keeps only rows where the name is "Elsa".
#.copy() makes copy of the filtered data.
elsa_df = df[df['name'] == 'Elsa'].copy()

#grouped/groupby turns it into a small df with two columns: year and total
elsa_grouped = elsa_df.groupby('year', as_index=False)['Total'].sum()

# Plot the trend with Frozen-style colors
(
    ggplot(elsa_grouped, aes(x='year', y='Total')) +
    geom_line(color='#39FF14', size=1.8) +  # light sky blue
    geom_smooth(se=False, method='loess', color='#4682B4') +  # steel blue smooth trend
    geom_vline(xintercept=2013, color='black', linetype='dashed', size=1.2) + 
    labs(
        title="Name 'Elsa' and the Impact of *Frozen*",
        subtitle="Dashed line marks the 2013 Disney release",
        x="Year",
        y="Total Babies Named Elsa",
    ) +
    #Creates a clean visual and removes unnecessary clutter
    theme_minimal()
)

Popularity of the Name Elliot Over Time

Looking at the graph it shows there wasn’t a huge popularity of the name Elliot until after a spike in 1982. This spike suggests that the popularity of the movie E.T. had a great influence on the popularity of the name Elliot. After the initial release there was a spike followed by a step decline. After the second release there was another increase and spike right around 1989 followed by a much more gradual decline. Howoever after ther third release there was a much more gradual incline in popularity. This continued popularity suggests that it can gain popularity on its own after being introduced.

Show the code

elliot_df = df[df['name'] == 'Elliot'].copy()

#groupeed/groupby puts year and sum total counts
elliot_grouped = elliot_df.groupby('year', as_index=False)['Total'].sum()

# Limits years 1950–2020
elliot_grouped = elliot_grouped[elliot_grouped['year'].between(1950, 2020)]

elliot_grouped['name'] = 'Elliot'

# Reference lines for E.T. movie releases
ref_lines = pd.DataFrame({
    'year': [1982, 1985, 2002],
    'label': ['E.T. Released', 'Second Release', 'Third Release'],
    'y': [1100, 1100, 1100] 
})
(
    ggplot(elliot_grouped, aes(x='year', y='Total', color='name')) +  # uses 'name' for legend title
    geom_line(size=.5) +  # purple/blue line
    scale_color_manual(values={"Elliot": "#6A5ACD"}, name="Name") + 

    # Vertical dashed red lines at key movie release years
    geom_vline(data=ref_lines, mapping=aes(xintercept='year'),
               color='red', linetype='dashed') +

    # Horizontal black text labels just above each red line
    geom_text(
        data=ref_lines,
        mapping=aes(x='year', y='y', label='label'),
        angle=0,
        hjust=0.8,
        vjust=-0.5,
        size=4.5,
        color='black',
        fontface='bold'
    ) +

    # Titles and axis labels
    labs(
        title="Elliot... What?",
        x="Year",
        y="Total",
    ) +

    # Lock axis limits
    scale_x_continuous(limits=[1950, 2020]) +
    scale_y_continuous(limits=[0, 1200]) +

    # Clean visual appearance
    theme_minimal()
)